Market moves: semis tick higher
The Philadelphia Semiconductor Index rose about 2.6% on strength tied to foundry signals, with NVIDIA up roughly 2.7% to $188.92 and AMD up about 5.1% to $248.66 in the same session — a fresh market nod to AI‑related demand. (x.com) Those moves also coincided with notable smart‑money inflows into memory and major AI ecosystem names, highlighting how differentiated capital is flowing into hardware plays. (x.com)
Chip stocks jumped again because investors got a fresh read on the one thing this market cares about most: whether the factories that build artificial intelligence chips are still running full. Taiwan Semiconductor Manufacturing, the main contract manufacturer for NVIDIA and Advanced Micro Devices, reported first-quarter 2026 revenue of NT$1.13 trillion, up 35% from a year earlier, and March revenue up 45%. (bloomberg.com) That matters because a foundry is the outside factory that actually makes chips designed by other companies. If Taiwan Semiconductor Manufacturing says orders are strong, traders read that as a live demand check on the whole artificial intelligence hardware chain. (nasdaq.com, bloomberg.com) The Philadelphia Semiconductor Index is the market’s scoreboard for 30 major chip names, and it closed at 8,889.69 on April 10, 2026, up 200.16 points, or 2.30%, on the day. CNBC also showed the index hitting a fresh 52-week high of 8,667.08 on April 9 before pushing higher again. (nasdaq.com, cnbc.com) NVIDIA rose to $188.63 on April 10, 2026, up 2.57% for the session, while Advanced Micro Devices closed at $245.04 on Yahoo Finance data after trading as high as $249.58 intraday on CNBC. Those are the two names most directly tied to the race to sell accelerators, which are the specialized chips used to train and run large artificial intelligence models. (cnbc.com, cnbc.com, finance.yahoo.com) The other piece of the trade is memory, which is the short-term working space those accelerators need to keep giant models moving. SK Hynix said in January that artificial intelligence demand drove its strongest quarterly results ever and that it expected a “considerable increase” in 2026 capital spending to expand capacity. (bloomberg.com) That is why money has been crowding into high-bandwidth memory, which is a stacked form of memory built to feed data to artificial intelligence chips much faster than ordinary server memory. SK Hynix said in its 2026 market outlook that demand for its high-bandwidth memory products would help drive a memory “supercycle” this year. (news.skhynix.com) Micron has become the United States market’s cleanest listed bet on that memory trade. On April 10, 2026, Micron closed with a market value of about $474.3 billion, and Wall Street coverage tied its rally to demand for high-bandwidth memory used in artificial intelligence systems. (cnbc.com, tradingkey.com) The reason the buying looks selective instead of broad is that not every chip company benefits the same way from artificial intelligence spending. Bloomberg reported in March that rising memory prices were already hurting some smartphone and personal computer demand even as Taiwan Semiconductor Manufacturing shifted capacity toward higher-end NVIDIA and Advanced Micro Devices data-center chips. (bloomberg.com) So the latest move was less a vote for “tech” than a vote for the parts of tech closest to data-center buildouts. NVIDIA, Advanced Micro Devices, Taiwan Semiconductor Manufacturing, Micron, and SK Hynix sit near the front of that line because they sell the compute, the factory capacity, or the memory that artificial intelligence servers cannot run without. (bloomberg.com, cnbc.com, bloomberg.com)
